What are autonomous services and why should businesses take notice? That was the primary focus of an in depth session with Oracle South Africa. The conversation turned out to be a great deal more, touching on how technology is set to shape the future, what businesses need to do to be competitive and relevant and the role that cloud, AI and big data will play.  

Samina Rizwan, the senior director of business analytics and big data for Oracle, began by laying out Oracle’s role in the cloud space. She explained that the company is 100% focused on cloud, with the distinction of being the largest scaled cloud company in the world, ahead of around 45 of the other players in the same space. “That is not to boast, but rather because our scale and size means that we can influence and provide services to a large portion of the market,” she clarified.

Autonomy ahead

As to what the company believes is the future, it’s autonomous services that are earning a good chunk of its focus these days. And for good reason. Rizwan explained that the poster child of autonomous services – and perhaps the easiest entry point to understand what autonomous services entail – is the self driving vehicle.    

Once the domain of science fiction, autonomous vehicles are quickly moving from the realm of possibility to actuality. Indeed, she conveyed that so confident is Oracle in the trend, that the company estimates that there will be 10 million self driving cars on the road across the world by 2020.

Rizwan explained that autonomous means that the product or service in question performs a function on its own. “In Oracle’s definition, for something to be autonomous, and be viable in the market, it must have three characteristics - it must be self-driving/operating, self-repairing (on a software level) and self-securing,” she elaborated. 

Intelligent enough?

She continued that, returning to the self driving car as an autonomous product, it needs to be self-securing, meaning that it has the right algorithms to recognize someone jaywalking, even if there isn’t a red traffic light, and stop accordingly.

If it cannot do that, then it’s not viable. However, if it has the right algorithms, based on artificial intelligence and machine learning, to emulate a human driver’s reasoning that a human crossing a road should be reacted to the same way as a stop sign or red traffic light, then it is self-securing and viable as an autonomous vehicle.

Indeed, Rizwan pointed out that the initial stages of self-driving cars resulted in accidents, where the car hit a pedestrian, came down to its programming, to save the passenger above all else, even if it meant colliding with an object. However, if that is augmented by programming that prioritises saving passengers and saving pedestrians, above having a collision, then it moves closer to an autonomous vehicle that can be trusted.

As for self-repairing, that has become a mainstay of Tesla vehicles, which download software updates and patches on their own, without needing to be taken into a service centre.

From a societal level, this means that AI and machine learning are critical components to having self driving cars that can be trusted and to which we would willingly surrender control.

Eradicating error

Beyond that, what about autonomous services and their role in IT and business?

Moving the example to the datacenter, Rizwan explained that autonomous services in a datacenter setting enable database administrator to do the necessary configuration in fifteen minutes, rather than three hours, as the machine learning parts of the technology take care of the mundane, time consuming tasks.

Additionally, integrating autonomous services into a data warehouse would allow it to install patches as needed on its own, identify a problem arising or a cybersecurity attack in the making and makes the necessary adjustments to fend it off without any external input being required.

“90% of crashes in data management environments happen because of human error, humans not noticing the signs of a problem. This, however, is completely resolved in an autonomous space,” she added. Indeed, modern computing is such that we just don’t expect the online services we use to go down, and are surprised when they do. However, Rizwan pointed out that we aren’t quite as shocked when we make a mistake, whether that entails losing our keys or burning our toast. Human error is an expected part of life and part of being infallible, but machine error is not.

In the second part, Rizwan looks at the speed of adoption that we can expect, at what autonomous services mean to businesses, and how global access is changing the game.